What is the Inception Network (GoogLeNet) and why is it important in deep learning? How does the Inception architecture improve feature extraction in convolutional neural networks? What is the role of inception modules in reducing computational cost while increasing performance? How does GoogLeNet differ from traditional CNN architectures? What are the real-world applications of Inception Networks in image recognition tasks?